FBI hair analysts gave erroneous testimony in 90%+ of cases over two decades, and 14 of the 32 defendants sentenced to death have been executed or died in prison

In 2015, the FBI acknowledged that 26 of 28 examiners in its microscopic hair comparison unit overstated forensic matches in ways that favored prosecutors in more than 95% of the 268 trials reviewed. The examiners testified that hair found at crime scenes 'matched' defendants' hair 'to a scientific certainty' or cited fabricated statistical probabilities like 'one chance in 10 million' that the hair could belong to someone else. No scientific basis exists for either claim. Microscopic hair comparison cannot identify a specific individual; at best it can say hair is 'consistent with' a source, which is a nearly meaningless statement given how similar human hair is across individuals. Of the 268 cases reviewed, 32 defendants were sentenced to death. Fourteen of those 32 have already been executed or died in prison. This is not a historical curiosity. The FBI's review has examined only a fraction of the cases in which hair testimony was used. The full review encompasses an estimated 2,500+ cases, and as of the most recent reporting, hundreds remain unreviewed. State and local labs that trained under FBI examiners replicated the same erroneous testimony in thousands of additional cases across the country, none of which are included in the FBI's count. Defendants convicted on hair evidence decades ago remain in prison today because post-conviction review processes are slow, underfunded, and require the defendant to initiate proceedings, often without legal representation. The 2009 National Academy of Sciences report labeled microscopic hair comparison 'highly unreliable,' but that finding does not automatically reopen any conviction. The structural reason this happened and continues to cause harm is that forensic disciplines were adopted by the criminal justice system based on practitioner claims rather than scientific validation. Hair examiners developed their own training programs, their own terminology, and their own standards of certainty, entirely outside the scientific peer-review process. Courts admitted this testimony for decades under the assumption that if an expert said it was science, it was science. The Daubert standard for expert testimony, which theoretically requires scientific validity, was not applied retroactively and is inconsistently enforced even in new cases. There is no mechanism to automatically reopen old convictions when the forensic method used to obtain them is later debunked.